Replacing a dying RAID1 drive on an array with a system?

Question

Iron

Greetings.
There is a server (file with a lot of access rights) with Intel Raid controller.
One of the disks in the Raid1 array, which is a system volume, assembled on SAS 15 K, judging by the Megaraid utility, began to die.
Symptoms:
- Backup works slowly;
- a lot of media errors in spec. utility;
- the rest of the arrays feel fine.
I ask for advice in the algorithm for replacing disks.

Answer 1

1. Backup everything
2. Pull out the disk
2.1 If it is possible to manually degrade the array on a broken disk
3. Install a new one similar
4. Run rebuild
5. Pray that the second disk is not covered
6. Similarly, replace the second disk (even if it is OK)

Answer 2

Practice on an identical controller
Make a backup of the system.
Replace disk.

Answer 3

Brief report on the results.
What did:
1. Turn off the server;
2. Replaced the failed drive with a new one (SAS 2.0 to SAS 3.0).
3. Launched the server, launched the rebuild (1 hour).
4. Turned off the server and repeated steps 2, 3, rebooted the server.
Bottom line:
The array is normal, but there is a suspicious feature - earlier in the MegaRaid utility (for old drives) the parameter "Drive Speed ​​\u003d 6.0 Gbps" was displayed, now "Drive Speed ​​\u003d UNKNOWN".